-
NEES/E-Defense Base-Isolation Tests: Effectiveness of Friction
Pendulum and Lead-Rubber Bearings Systems
Tomohiro SASAKI & Eiji SATO National Research Institute for
Earth Science and Disaster Prevention
Keri L. RYAN University of Nevada, Reno
Taichiro OKAZAKI Hokkaido University
Stephen A. MAHIN University of California, Berkeley
Koichi KAJIWARA National Research Institute for Earth
Science and Disaster Prevention
SUMMARY: This paper reports on a Network for Earthquake
Engineering Simulation (NEES)/E-Defense collaborative tes t ing
program on base-isolated buildings. A full-scale, five-story,
two-by-two bay, steel moment-frame building was subjected to a
number of bidirectional and bidirectional-plus-vertical ground
motions using the E-Defense shake table. The building was tested
under three different configurations: 1) base isolated with
triple-friction-pendulum bearings (TPB), 2) base isolated with a
combination of lead-rubber bearings (LRB) and cross-linear bearings
(CLB), and 3) fixed-base. This paper introduces the results of
shake table experiments on full-scale base-isolated building and
shows the effectiveness of recent base-isolation techniques for
frequent, near-fault and long duration subduction earthquakes.
Based on experimental results, it was found that TPB system
provided greater attenuation of floor accelerations for the ground
motions with PGA larger than 10 m/s2 while LRB-CLB system provided
greater attenuation of floor accelerations for ground motions with
PGA smaller than 5 m/s2.
Keywords: seismic design, base-isolation, triple friction
pendulum bearing, lead-rubber bearing, steel moment frame
building
1. INTRODUCTION
This paper reports on a collaborative testing program on
base-isolated buildings conducted under the Memorandum of
Understanding between the National Research Institute for Earth
Science and Disaster Prevention (NIED) of Japan and the National
Science Foundation (NSF), George Brown Jr. Network for Earthquake
Engineering Simulation (NEES) program of the U.S.
Base isolation is one of the most effective measures to
protect building structures and their nonstructural components from
earthquake ground motions. Development of modern seismic isolation
technique started in the 1960’s in New Zealand (Skinner et al.,
1993). Various forms of elastomeric bearings such as natural rubber
bearings, high damping rubber bearings, and lead-rubber bearings
have been commercially implemented since the 1980’s. The friction
pendulum bearing is a newer base isolation device that was first
developed in the late 1980’s (Zayas et al., 1987) and whose
implementation to buildings started from a seismic retrofit project
in 1994. Currently, a wide range of base isolation devices are
commercially available and are being implemented in practice.
However, wide acceptance of base isolation has not happened yet
partly because their cost benefit is not well understood in the
structural engineering community.
Consequently, a large-scale, shake-table test program was
conducted with a goal of promoting rapid spread of base isolation
systems in Japan and the U.S. In this program, a full-scale,
five-story, steel
-
moment-frame building was subjected to a number of bidirectional
and bidirectional-plus-vertical ground motions using the world’s
largest shake table, E-Defense. The building was tested under three
different configurations: 1) base isolated with
triple-friction-pendulum bearings (TPB), 2) base isolated with a
combination of lead-rubber bearings (LRB) and cross-linear bearings
(CLB), and 3) base fixed. This paper compares the response obtained
for the three configurations and discusses the effectiveness of
recent base-isolation techniques to protect buildings, along with
its contents and nonstructural elements, from frequent, near-fault,
and long duration subduction earthquakes.
2. SPECIMEN AND EXCITATION
2.1. Steel Moment-Frame Building
Fig. 1 shows the superstructure used in the shake table tests.
The superstructure was a five-story, two bay-by-two bay steel
moment-frame building that was designed and tested in a previous
program by Kasai et al. (2010). While dampers were placed in four
of the five stories in the previous program, the dampers were
removed from the building for this program. As shown in Fig. 1, X
and Y directions were defined as the direction of short and long
sides of the slabs and Z direction was defined as the vertical
direction. The building was 15.8-m tall and had a floor plan
dimension of 10 m by 12 m. The columns used 400-mm deep square
hollow sections with thickness ranging between 12 and 22 mm. The
girders and beams used wide flange sections with dimension of 400
mm × 200 mm, 300 mm × 150 mm, and 198 mm × 99 mm. The base beams
were relatively rigid wide flange beams with dimension of 900 mm ×
500 mm. All girder-to-column and column base connections were
fully-restrained moment connections. Each girder had short end
segments welded to the column in shop, and the remaining middle
segment of the girder was spliced in the field by bolting. The
girder flanges were widened near the connections to the column to
avoid fracture at the critical welds. At the second to fifth
floors, composite slabs were formed from 75-mm height corrugated
steel decks covered by 80-mm thick normal concrete. The roof slab
was a 150 mm thick normal concrete slab with a flat steel deck. The
total weight of the building was 486 tons. A 56-ton steel mass was
added on the east side and middle of the roof to represent a
penthouse and to intentionally introduce eccentricity in mass and
weight.
Nonstructural components were installed on the fourth and
fifth floors. U.S.-style suspended ceiling grid systems, interior
walls, and piping systems were installed using U.S. material and
construction techniques. The architectural layout included enclosed
rooms to allow for the enactment and portrayal of hospital and
office rooms including a variety of furniture and other loose
items. In addition, two full story pre-cast concrete cladding
panels were set between the fourth and fifth floors to evaluate the
effectiveness of current slotted steel connection design to allow
story drift. Details of the nonstructural
Figure 1. Five-story Steel Moment Frame Building
-
(a) Triple Friction Pendulum Bearing (TPB)
(b) Lead-Rubber Bearing (LRB) (c) Cross Linear Bearing (CLB)
Figure 2. Isolation Devices
components and their behavior are reported in a companion paper
by Soroushian et al. (2012).
2.2. Base-Isolation Devices
Two different base isolation systems were devised for the shake
table program. The first system placed nine triple friction
pendulum bearings (TPB) under each of the nine columns. The second
system was a combination of four lead-rubber bearings (LRB) and
five cross linear bearings (CLB). Fig. 2 shows isolation devices
used in this study. After testing of the two base-isolation systems
was completed, the superstructure was directly tied to the shake
table and tested under the base-fixed condition.
The TPB system was designed to meet specific performance
objectives for U.S. MCE and Japan L3 event. U.S. MCE spectrum was
computed from spectral coefficients SMS = 2.2g and SM1 = 1.11g at
0.2 and 1.0 seconds, respectively, assuming a Los Angeles location
with site class D (ASCE 2010). The Japan L3 spectrum was computed
as a 50% increase over an L2 event at the site with the soil type
II in Zone A seismicity, which applies to most regions of Japan
(AIJ, 2001 and Pan et al., 2005). The design parameters of TPB such
as friction coefficients, effective pendulum lengths and
displacement capacities of the three independent pendulum
mechanisms were selected to achieve the target demands of floor
accelerations < 0.35g and story drifts < 0.5%. The TPBs were
1.4 m in diameter, 0.33-m tall, with displacement capacity of 1.13
m. The natural periods were 1.84 and 5.57 seconds, and friction
coefficients were 0.02 and 0.08 in the inner and outer pendulum
mechanisms, respectively.
The LRB-CLB system placed LRBs under the four edge columns
and CLBs under the four corner and one center column, considering
the relatively light weight of the steel moment-frame building. The
LRB-CLB system was designed to protect a nuclear power plant in
beyond-design-basis shaking. In this project, design-basis shaking
was determined as ground motions with an annual probability of
exceedance of 10-5 for a representative eastern U.S. soil site,
where the design spectrum was determined from a site specific
analysis (Huang et al. 2009). The LRBs were 0.699 m in diameter and
0.46-m tall. Assuming a constant axial force of 1,325 kN per
bearing based on the tributary gravity load, these LRBs provide a
yield strength coefficient of V/W = 0.055 and a post-yield natural
period of 2.86 sec. The CLBs were 0.448 m tall and comprised two
1.65-m long rails in two orthogonal directions. Each rail had a
displacement limit of ± 0.6 m. The CLBs were rated to provide a low
friction coefficient of 0.0019 under a constant vertical
compression of 485 kN.
2.3. Instrumentation
More than 650 sensor channels were used to measure the
response of the isolation system, the superstructure, and the
nonstructural components. For the isolation system, wire
potentiometers were used to measure the bidirectional displacements
and an assembly of tri-axial load cells was installed under each
bearing to measure the forces. For the superstructure,
accelerometers were placed on three
-
corner columns on each floor. Additional accelerometers were
placed on the floor slab to measure floor vibration. Laser-type
displacement transducers were used to measure the story drift of
each story. More than 200 channels were devoted to nonstructural
components. Accelerometers and displacement transducers were placed
on ceiling, interior walls, piping, and exterior pre-cast walls to
measure their dynamic response. 2.4. Excitation Plan
The building was subjected to various ground motions in each
of the three configurations. The shake table tests were conducted
over six days; three days for the TPB system, two days for the
LRB-CLB system, and one day for the fixed-base case. The motions
for the TPB system were selected with the objective of subjecting
the system to a wide variety of strong ground motions with
different characteristics, ranging from intense high frequency
content (e.g. JMA Kobe record from the 1995 Kobe earthquake), near
fault (e.g. Sylmar record from the 1994 Northridge earthquake),
very long period (e.g. ChiChi TCU065 record from the 1999 ChiChi
earthquake), to long-duration subduction (e.g. Iwanuma record from
the 2011 Tohoku earthquake). The motions for LRB-CLB system were
selected primarily with the objective of approaching the
displacement limit of the system using synthetic motions
representative of potential nuclear sites that had been studied by
Huang et al. (2009) (i.e., Vogtle on the U.S. east coast and Diablo
Canyon on the west coast). The motions for the fixed-base case were
selected with the objective of establishing a base of comparison
between isolated and fixed-base buildings. The motions for the
fixed-base case were scaled down to maintain the structure response
within the linear range. In addition to the primary excitations,
white noise and sweep excitations were conducted to enable system
identification in the fixed-base configuration. This paper
uses the data obtained from three motions, Westmorland, Iwanuma,
and Rinaldi records, to compare the response of the building under
the different configurations. All three are recorded motions. Fig.
4 shows the time history of the three target motions, while Fig. 5
shows the response acceleration and displacement spectra. For
reference, Fig. 5 indicates the Japan L1, L2, and L3 design
spectrum. The Westmorland motion was obtained at Westmorland from
the 1987 Superstition Hills earthquake. For this program the
original motion was scaled down to 80% for all three
configurations. The Iwanuma motion was obtained at Iwanuma, Miyagi,
from the 2011 Tohoku earthquake. The 100% horizontal motion was
used for the TPB and LRB-CLB systems and the 70% horizontal
motion
-
Res
pons
e A
ccel
erat
ion
(m/s
) Tr
ansf
er F
unct
ion
A/A
R
espo
nse
Acc
eler
atio
n (m
/s)
Res
pons
e A
ccel
erat
ion
(m/s
)
Res
pons
e D
ispl
acem
ent
(x10
2 m
m)
Res
pons
e D
ispl
acem
ent
(x10
2 m
m)
Res
pons
e D
ispl
acem
ent
(x10
2 m
m)
Tran
sfer
Fun
ctio
n A
/A
7 7 6 X 6 X Z 5 Y 5 Y L1 4 Z 4 3 L1 3 2 2 1 1 0 0 0 2 4 6 8
10
Period (sec) 0 2 4 6 8 10
Period (sec) (a) 80% Westmorland motion
X 6
15 Y 5
10 L2 4
3 5 2
1 0 0
Period (sec)
X L2 Y
Period (sec)
(b) 100% Iwanuma motion 20 7
X 6 15 Y 5
10 L3 4 3 5 2
1 0 0
X L3 Y
0 2 4 6 8 10 Period (sec)
0 2 4 6 8 10 Period (sec)
(c) 88% Rinaldi motion Figure 5. Response Spectrum
was used for the fixed-base case. The Rinaldi motion was
the obtained at Rinaldi Receiving Station from the 1994 Northridge
earthquake. The Rinaldi motion is characterized by extremely strong
vertical components. An 88% horizontal-only motion was used for the
TPB and LRB-CLB systems, while a 35% horizontal-only motion was
used for the fixed-base case. The response to 3D Rinaldi
excitation, which was another motion used for all three
configurations, is the subject of a companion paper by Ryan et al.
(2012) and is not discussed in this paper.
3. SYSTEM IDENTIFICATION OF SUPERSTRUCTURE
Fig. 6 shows the transfer functions determined from the white
noise excitation conducted for the fixed-base case prior to the
primary excitations. The period and damping ratio corresponding to
the fundamental response modes were evaluated by curve fitting
theoretical transfer functions to the measured transfer functions
using a least square algorithm. It should be noted that, because
rocking motion of the table affected the system identification
results significantly, Eqs. 19 and 20 in Kasai et al. (2011) were
used to eliminate the rocking motion. Table 1 and Fig. 7 summarize
the evaluated periods
12 12 Experiment
10 Fitting Curve 10
8 8
6 6
4 4
2 2
0 0
Experiment Fitting Curve
0 2 4 6 8 10 Frequency (Hz)
0 2 4 6 8 10 Frequency (Hz)
Figure 6. Transfer Function during Whitenoise Excitation before
Test in fixed-base configuration
-
Dam
ping
Rat
io
Table 1. Natural Periods and Damping Ratios (a) X direction (b)
Y direction
Period (sec) Damping Ratio Period (sec) Damping Ratio 1st 0.718
0.0643 1st 0.702 0.0668 2nd 0.454 0.0946 2nd 0.452 0.0545 3rd 0.207
0.0223 3rd 0.212 0.0224 4th 0.149 0.0355 4th 0.113 0.0315 5th 0.112
0.0363 5th 0.073 0.0342 6th 0.074 0.0239 6th 0.055 0.0185
0.15
0.1
0.05
X Y Rayleigh (αm+βk)
α = 1.39 β = 3.71 x 10-4
0 0 2 4 6 8 10
Frequency (Hz) Figure 7. Damping Ratio
and damping ratios. As indicated in Fig. 7, the damping
ratios can be reasonably approximated by Rayleigh damping with
coefficients α =1.39 and β = 3.71 ×10 −4 .
4. EXPERIMENTAL RESULTS
4.1. Isolator Response
Fig. 8 shows the bidirectional displacement and twist angle
obtained from the TPB and LRB-CLB systems for each of the three
motions. For the 80% Westmorland motion shown in Fig. 8(a), the
horizontal displacements in the two systems were surprisingly
similar to each other considering the difference in their effective
natural periods. For the 100% Iwanuma motion shown in Fig. 8(b),
the peak isolator displacements were very similar, reaching 368 mm
in the TPB system and 347 mm in the LRB-CLB system. On the other
hand, the twist angle was substantially larger in the LRB-CLB
system than in the TPB system. For the 88% Rinaldi motion shown in
Fig. 8(c), the peak twist angle in the LRB-CLB system was 0.96 o ,
which corresponds to a 201-mm difference in X-displacement between
the east and west edges. In the TPB system, twist angle was 0.19 o
, which corresponds to a mere 4-mm difference in X-displacement
between the two edges. Because the horizontal reaction of the TPB
is proportional to the vertical force (Zayas et al., 1987), the
center of gravity always coincides with the center of reaction in
TPB systems.
Fig. 9 shows the hysteresis of the isolation layer for
each of the three motions, plotting the base shear against
isolation layer displacement. For the 80% Westmorland motion shown
in Fig. 9(a), the hysteretic loop was very similar between the TPB
system and LRB-CLB system. It was observed in the TPB system that,
because the response displacement was small such that sliding
occurred only along the inner sliding surface and the outer sliding
was not activated, the natural period was close to the effective
period of the LRB-CLB system. For the 100% horizontal-only Iwanuma
motion shown in Fig. 9(b), the peak base shear reached 1,187 kN (=
0.224W) in LRB-CLB system and 719 kN (= 0.135W) in TPB system. For
the Rinaldi motion shown in Fig. 9(c), the TPB system exhibited a
very uneven hysteretic loop. As discussed further in a companion
paper by Okazaki et al. (2012), the uneven hysteresis was due to
substantial uplift in several TPBs.
-
Base
She
ar (k
N)
Bas
e Sh
ear (
kN)
Bas
e Sh
ear (
kN)
800 600 X Y 400 200
0 -200 -400 -600 -800
TPB LRB
TPB LRB
-100 -50 0 50 100-100 -50 0 50 100 Displacement (mm)
Displacement (mm)
1500
(a) 80% Westmorland motion 1500
1000 X Y 500
1000 X Y 500
0 -500 -1000 TPB
LRB -1500
TPB LRB
0 -500
-1000 TPB LRB
-1500
TPB LRB
-400 -200 0 200 400-400 -200 0 200 400 -400 -200 0 200 400-400
-200 0 200 400 Displacement (mm) Displacement (mm) Displacement
(mm) Displacement (mm)
(b) 100% Iwanuma motion (c) 88% Rinaldi motion Figure 9. Base
Shear vs. Isolation Layer Displacement Hystereses
4.2. Floor Accelerations Figs. 10 and 11 shows the
floor accelerations in the Y direction measured from the 100%
Iwanuma and 88% Rinaldi motions. The acceleration histories were
taken from the first and fifth floors for the TPB and LRB-CLB
systems and from the fifth floor for the fixed-base case. The floor
acceleration was substantially smaller in the TPB system than in
the fixed-base case. While the same can be said for the LRB-CLB
system, this system experienced several impulse accelerations at
the first floor. The impulse was caused by slipping of the bolts at
the top and bottom connections of the LRBs. Due to experimental
constraints, a minimal number of bolts were used for these
connections. However, these
-
impulses did not cause damage to the superstructure.
Because the impulse motion was not transmitted to the upper floors,
the impulse did not cause damage to the loose contents or the
nonstructural elements. In conclusion, Figs. 10 and 11 suggest that
both isolation systems worked effectively and as intended.
Fig. 12 plots the ratio of root-mean-square (RMS) between the
floor acceleration aiF ,rms and table acceleration a g ,rms ,
evaluated for all floors. The fixed-base case experienced response
amplification in the form of aiF ,rms / a g ,rms exceeding unity
and increasing with floor height. However, the aF,rms/ag,rms value
was below unity in both isolation systems, indicating that both
isolation systems worked as intended for frequent (80%
Westmorland), long duration subduction (100% Iwanuma), and near
fault (88% Rinaldi) ground motions. The amplification factor
aF,rms/ag,rms measured from the 80% Westmorland motion (Fig. 12(a))
was 0.84 to 0.95 for the TPB system and 0.73 to 0.92 for the
LRB-CLB system. Interestingly, the factor measured from the 88%
Rinaldi motion (Fig. 12(c)) was smaller at 0.33 to 0.44 for the TPB
system and 0.42 to 0.56 for the LRB-CLB system. The peak measured
table acceleration was 1.49-1.53 m/s2 and 11.25-12.03 m/s2 for the
Westmorland and Rinaldi motion, respectively. Therefore, the
-
5F,rm
s g
,rms
RF 10
5F
4F 1
3F
2F 0.1
TPB LRB Fix
TPB TPB TPB 0 5 10 15 20 1F LRB LRB LRB PGA a
g (m/s2)
Table Fix 0 1 2 3 4 Fix
0 1 2 3 4 Fix
0 1 2 3 4 Figure 13. Dependence of Acceleration Amplification
Factor by
a iF,rms
/ a g,rms
a iF,rms
/ a g,rms
a iF,rms
/ a g,rms
Isolation Devices on Peak (a) Westmorland motion (b)
Iwanuma motion (c) Rinaldi motion
Figure 12. Ratio of RMS between Response Acceleration and Table
Acceleration (Y direction)
Table Acceleration (Fifth floor)
comparison indicates that both isolation systems were more
effective for motions with larger PGAs. Also, the TPB system
provided greater attenuation of floor acceleration than the LRB-CLB
system against motions with larger PGA, during which the TPB system
response was dominated by sliding on the outer pendulum surfaces,
with an isolation period of 5.57 sec compared to 2.86 seconds for
LRB-CLB. However, the LRB-CLB system provided greater attenuation
of floor acceleration against motions with smaller PGA, during
which the TPB system response was dominated by sliding on the inner
sliders, with an isolation period of only 1.84 seconds. These
observations are a result of the isolation periods provided by the
systems, and not the mechanics of the bearing devices.
Fig. 13 summarizes the relationship between the
amplification factor at the fifth floor a5 F ,rms / a g ,rms and
PGA computed for all motions imposed to the specimen expect for 3D
Rinaldi motion containing strong vertical acceleration. As observed
from Fig. 12, the LRB-CLB system tended to be more effective than
the TPB system when PGA was smaller than 5 m/s2. Recall that the
TPBs used for the specimen were designed to respond rigidly until
the horizontal force overcomes the friction force between the inner
sliding surfaces, 0.02W. However, when the PGA exceeded 10 m/s2,
the TPB system tended to be more effective than the LRB-CLB system
in reducing the amplification factor.
4.3. Damage to Nonstructural Components
The photographs in Fig. 14 show the damage to
nonstructural components observed for the fixed-base case. During
the 35% horizontal-only Rinaldi motion, contents moved violently
and items fell from shelves. Damage to the ceiling system was
limited to minor cracking around the sprinkler heads. During the
70% horizontal-only Iwanuma motion, the ceiling system suffered
damage such as falling of panels and bending of the metal members
comprising the grid. After testing, the tables and shelves had
moved and almost all loose items had dropped to the floor. In
contrast, for the TPB and LRB-CLB systems, no damage was observed
in the ceiling systems, interior walls, or piping. Movable contents
such as chairs with wheels rolled around, but otherwise, no damage
was observed in the contents.
5. CONCLUSIONS
A large-scale shake-table test program was conducted under the
collaborative research program between E-Defense and NEES with a
goal of promoting rapid spread of base-isolation systems in Japan
and the U.S. Based on the experimental results, the following
conclusions are deduced;
-
1) In both the TPB and LRB-CLB isolation systems, the peak floor
accelerations were smaller to those measured in fixed-base case.
Both isolation systems successfully mitigated damage in the
superstructure from a range of strong ground motions. The
amplification factor, which is the ratio of root-mean-square values
of the floor acceleration and table acceleration aF,rms/ag,rms,
measured from the 80% Westmorland motion with PGA of 1.49-1.53 m/s2
was 0.84 to 0.95 for the TPB system and 0.73 to 0.92 for the
LRB-CLB system. The amplification factor measured from the 88%
Rinaldi motion with PGA of 11.25-12.03 m/s2 was smaller at 0.33 to
0.44 for the TPB system and 0.42 to 0.56 for the LRB-CLB system. In
both systems, the amplification factor decreased with increasing
PGA. 2) TPB system provided greater attenuation of floor
accelerations for the ground motions with PGA larger than 10 m/s2
while LRB-CLB system provided greater attenuation of floor
accelerations for ground motions with PGA smaller than 5 m/s2.
These observations are a result of the isolation periods provided
by the systems. 3) Due to the intentional mass eccentricity, the
isolation layer was subjected to substantial torsion. While the two
systems were not specifically designed for torsion, twisting motion
was more effectively mitigated in the TPB system than in the
LRB-CLB system.
ACKNOWLEDGEMENT Funding for this study was provided by NIED, the
National Science Foundation through Grants No. CMMI-1113275 and
CMMI-0721399, and U.S. Nuclear Regulatory Commission through
Contract NRC-HQ-11-C-04-0067. The isolation devices and design
service was donated by Earthquake Protection Systems, Dynamic
Isolation Systems, and Aseismic Devices Company. The views
reflected in this paper are those of the authors alone and do not
necessarily reflect those of the sponsors.
REFERENCES
American Society of Civil Engineers (2010). Minimum Design Loads
for Buildings and Other Structures, ASCE Standard ASCE/SEI 7-10,
Reston, VA, USA.
Architectural Institute of Japan (2001). Recommendation for the
Design of Base Isolated Structures (in Japanese).
Huang, N. Y., Whittaker, A. S., Kennedy, R. P. and Mayes, R. L.
(2009). Assessment of Base-Isolated Nuclear Structures for Design
and Beyond-Design Basis Earthquake Shaking, Tech. Report
MCEER-09-0008, University at Baffalo, USA.
Kasai, K., Ito, H., Ooki, Y., Hikino, T., Kajiwara, K., Motoyui,
S., Ozaki, H. and Ishii, M. (2010). Full scale shake table tests of
5-story steel building with various dampers, Proc. 7th Intern.
Conf. on Urban Earthquake Engin. & 5th Inern. Conf. on
Earthquake Engin., pp. 11-22, Tokyo Institute of Technology, Tokyo,
Japan.
Kasai, K., Murata, S., Kato, F., Hikino, T. and Ooki, Y. (2011).
Evaluation Rule for Vibration Period, Damping, and Mode Vector of
Buildings Tested by a Shake Table with Inevitable Rocking Motions,
J. Struct. Constr. Eng., AIJ, Vol. 76, No. 670, pp. 2031-2040 (in
Japanese).
Okazaki, T., Sato, E., Ryan, K. L., Sasaki, T., Mahin, S. and
Kajiwara, K. (2012). NEES/E-Defense Base Isolation Tests: Peromance
of Bearings, Proc. 15th World Conf. on Earthquake Engin., Lisbon,
Portugal.
Ryan, K. L., Dao, N. D., Sato, E., Sasaki, T. and Okazaki, T.
(2012). NEES/E-Defense Base Isolation Tests: Interaction of Lateral
and Vertical Response, Proc. 15th World Conf. on Earthquake Engin.,
Lisbon, Portugal.
Skinner, R. I., Robinson, W. H. and McVerry, G. H. (1993). An
Introduction to Seismic Isolation, John Wiley & Sons Inc., NY,
USA.
Soroushian, S., Ryan, K. L., Maragakis, M., Sato, E., Sasaki,
T., Okazaki, T., Tedesco, L., Zaghi, A. E., Mosqueda, G. and
Alvarez, D. (2012). NEES/E-Defense Tests: Seismic Performance of
Ceiling / Sprinkler Piping Nonstructural Systems in Base Isolated
and Fixed Base Building, Proc. 15th World Conf. on Earthquake
Engin., Lisbon, Portugal.
Zayas, V., Low, S. and Mahin, S. (1987). The FPS Earthquake
Resisting System, Technical Report UCB/EERC-87/01, Earthquake
Engineering Research Center, University of California, Berkeley,
CA, USA.
